INDEX

Explanations

references to something being empty or vacant

oai_token-act-pair · gpt-3.5-turbo

instances of the word "empty" or its variations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arya

-0.83

iane

-0.72

abol

-0.69

Downloadha

-0.69

ovan

-0.67

ocard

-0.66

multi

-0.66

ection

-0.65

ector

-0.63

ect

-0.63

POSITIVE LOGITS

 shelves

0.91

 spaces

0.86

 bottles

0.86

 space

0.85

 calories

0.84

 shells

0.81

 stomach

0.81

hus

0.81

 storefront

0.79

 cavity

0.77

Activations Density 0.050%

references to something being empty or vacant

instances of the word "empty" or its variations

No Comments

No Known Activations